Retinal signals are relayed to cerebral cortex through the lateral geniculate nucleus (LGN). The LGN is divided into several layers, in part segregating these signals by eye and functional class of retinal ganglion cell. In the cat, different layers also vary in the degree to which individual relay cells summate the output of multiple retinal cells, and thus the degree to which they gain sensitivity at the expense of spatial resolution. Long term objectives of this project are to understand how and why the LGN is organized into layers, what these channels contribute to perception and visuomotor behavior, how the information relayed through them is integrated in cerebral cortex, and how central feedback to the LGN dynamically regulates transmission of information to cortex. Contributions of individual layers to visuomotor behavior will be examined by briefly inactivating selected layers in cats carrying out oculomotor tasks. To test the hypothesis that LGN subdivisions vary in the degree to which they are specialized for high-acuity or high-sensitivity vision, inactivations will be done under conditions emphasizing the importance of either spatial resolution or dim-light vision. A similar paradigm will be used to test the hypothesis that the relative influence of different LGN layers on cortical activity varies depending on task demands and adaptations levels. Permanent lesions of dorsal LGN layers will be combined with reversible inactivations of central layers to examine the nature of a recovery of function that appears to occur that such lesions. This novel form of plasticity may shed light on perceptual deficits resulting from abnormalities in particular afferent saccadic eye movements. Because the purpose of saccades is to move the target of interest to central retina regions of the brain representing central retina might be subject to unique forms of saccade-elated changes in gain. This possibility will be examined for cells in the LGN representing the area centralis. To better understanding developmental rules guiding LGN morphogenesis, the complex lamination pattern of the human LGN will be visualized thorough 3-D computer reconstructions of individual nuclei to determine if a computational model of morphogenesis developed for the rhesus monkey can be applied to the human.